In conventional spectral CT imaging, it has been shown that energy-weighting, whether in the image-domain or the data-domain, improves the contrast-to-noise ratio. Conventionally, photon-counting detectors use a set of predetermined energy thresholds to count the number of events falling into predetermined energy ranges. However, no practical implementation of energy-weighting at the detector level is known.